critical clearing time and wind power in small isolated

Article

Critical Clearing Time and Wind Power in SmallIsolated Power Systems ConsideringInertia EmulationElías Jesús Medina-Domínguez 1,*,: and José F. Medina-Padrón 2,:

Received: 5 August 2015 ; Accepted: 2 November 2015 ; Published: 6 November 2015Academic Editor: Frede Blaabjerg

1 Renewable Energies Deparment, Research and Development Division, Canary Islands Institute ofTechnology (ITC), C/Playa de Pozo Izquierdo s/n, Santa Lucía (Gran Canaria) 35119, Spain

2 University Institute of Intelligent Systems and Numeric Applications in Engineering (SIANI),University of Las Palmas de Gran Canaria (ULPGC), Edificio Central del Parque Científico y Tecnológico,Campus Universitario de Tafira, Las Palmas de Gran Canaria 35017, Spain; [email protected]

* Correspondence: [email protected]; Tel.: +34-928-727-560: These authors contributed equally to this work.

Abstract: The stability and security of small and isolated power systems can be compromised whenlarge amounts of wind power enter them. Wind power integration depends on such factors aspower generation capacity, conventional generation technology or grid topology. Another issuethat can be considered is critical clearing time (CCT). In this paper, wind power and CCT arestudied in a small isolated power system. Two types of wind turbines are considered: a squirrelcage induction generator (SCIG) and a full converter. Moreover, the full converter wind turbine’sinertia emulation capability is considered, and its impact on CCT is discussed. Voltage is taken intoaccount because of its importance in power systems of this kind. The study focuses on the small,isolated Lanzarote-Fuerteventura power system, which is expected to be in operation by 2020.

Keywords: wind power; isolated power system; transient stability; full converter wind turbine;inertia emulation

1. Introduction

The unprecedented worldwide development of wind power over recent decades reached around318,644 GW by the end of 2014, and it is estimated that global wind power capacity will continue toincrease [1].

Integrating wind power into power systems is an increasingly common challenge. Largeamounts of wind power can be introduced into continental power systems, such as the European gridmanaged by the members of the European Network of Transmission System Operators for Electricity(ENTSO-E). On the other hand, small, isolated power systems, such as those found on small islands,can experience problems related to stability, and this can influence wind integration [2].

One issue that deserves consideration in wind power integration is the critical clearing time(CCT), particularly in small, isolated power systems. CCT can be seen as a measure of the transientstability of a power system [3,4] and can be defined as the maximum allowed duration of athree-phase short circuit before loss of system stability. As a part of the Spanish Grid Codes, a setof protection coordination criteria and a methodology leading to CCT evaluation were developedby the Spanish Transmission System Operator (TSO) Red Eléctrica de España (REE) [5,6]. The CCTcalculation methodology implies performing dynamic stability simulations by applying a short circuitin the nodes of the network and varying the fault time until a set of criteria are fulfilled [7,8]. The

Energies 2015, 8, 12669–12684; doi:10.3390/en81112334 www.mdpi.com/journal/energies

Energies 2015, 8, 12669–12684

set of criteria to be fulfilled through this trial and the error approach specify that the following isnot allowed:

(1) loss of system stability;(2) an unacceptable value of load shedding;(3) unacceptable parameters values in the subsequent steady state.

The Spanish TSO has established 10% of load as the unacceptable value for Spanish powersystems on islands and in other small systems [6].

There are other methods for CCT estimation, such as Lyapunov energy function, artificialneuronal networks or hybrid methods. However, the trial and error method using dynamicsimulations provides more accuracy results. In this paper, we use the trial and error method andthe above-mentioned criteria to study how using inertia emulation could impact CCT values andeventually the wind power penetration level.

A TSO uses CCT for designing protection schemes in power systems. A significant modificationof the CCT values can cause problems in protection system behavior that can have an impact on thestability or security of the power system itself.

Several papers and reports have investigated wind power and system stability through CCT,and their results show that there is a relationship between them. They suggest a modification of CCTvalues when wind power is introduced into a power system [9–11].

Other papers have analyzed inertia emulation from wind turbines and have shown a positiveimpact of the inertia emulation on the frequency response [12–20].

These papers usually look at large power systems and conventional generation based on largethermal or hydro generator units. Generation unit outages or transmission tie line outages areexamples of the types of disturbances studied in papers considering inertia emulation. Medium andlow voltage networks or feeders are also studied, and the analysis in these papers focuses on the timeevolution of frequency.

In [21], the power system of an island and wind farms with inertia emulation are considered.The integration of a hydro-pump storage system was analyzed, but short-circuit disturbance was nottaken into account.

In this paper, the relationship between wind power and CCT is studied in a small, isolatedpower system with two types of wind turbines: the squirrel cage induction generator (SCIG) andthe full converter. Furthermore, the inertia emulation capability of the full converter wind turbineis considered. A comparative analysis of the different results achieved is presented. In order toobtain CCT values, a three-phase short circuit disturbance is simulated. Therefore, inertia emulationperformance was studied when a three-phase short circuit takes place. Network voltages are alsoconsidered because of their relevance in this type of power system. Thus, a more global analysisof the small and isolated power system can be undertaken. Almost all of the system’s conventionalgeneration is based on small diesel generators.

The study has been carried out on the planned Lanzarote-Fuerteventura power system, which isexpected to be in operation by 2020. Its generation power capacity, voltage level, inertia constant andgrid topology are features that make it a good example of a small, isolated power system.

Modeling and analysis are performed using PSSrE v32 software (Siemens-PTI, Schenectady,NY, USA).

2. Methodology

CCT values are obtained through a trial and error procedure [22] when using dynamic analysisfor several three-phase short circuit events at some Lanzarote-Fuerteventura power system buses.Simulation results were studied to find out what determines the CCT.

These short circuit disturbances were simulated at five selected buses in order to know theoverall behavior of the power system. The selected buses are: (1) Punta Grande power plant and

12670

Energies 2015, 8, 12669–12684

Las Salinas power plant; (2) Haría-Teguise and Jandía as the farthest buses from power plants; and(3) Corralejo, which is a bus in between both power plants.

In order to analyze the wind power impact on CCT, three types of wind turbine were studied:SCIG, full converter and full converter with inertia emulation capability. Each wind farm in thepower system is equipped with only one kind of wind turbine. Thus, their CCT values can showsome differences.

In every case, wind power was increased from 0 MW to 150 MW, in steps of 10 MW. This windpower was generated by wind farms according to their rated powers.

In order to find CCT values, the CCT definition given in the Introduction section is used. In thisway, the unacceptable value of load shedding is equal to or higher than 10% of the load.

Photovoltaic power plants were not modeled to avoid their effects on the results.The model of the Lanzarote-Fuerteventura power system for steady-state analysis and dynamic

simulation was created using PSSrE v32.

3. Description of the Studied Power System

The system studied here is the Lanzarote-Fuerteventura power system 2020. Lanzarote andFuerteventura are two of the Spanish Canary Islands.

The Lanzarote and Fuerteventura power systems are linked by a submarine cable and, therefore,constitute a single power system.

At present, each island has only one conventional power plant. On Lanzarote, Punta Grande hasa capacity of 232.4 MW, and Las Salinas, on Fuerteventura, has a capacity of 187.0 MW. Both powerplants have conventional diesel and gas turbine units. New power generators are not expected to beinstalled in the foreseeable future, because power demand has decreased over recent years [23]. Theunits at these power plants are listed in Table 1.

Table 1. Conventional generation units.

Punta Grande unit name Capacity (MVA) Las Salinas unit name Capacity (MVA)

Diesel 1 9.4 Diesel 1 5.4Diesel 2 9.4 Diesel 2 5.4Diesel 3 9.4 Diesel 3 6.3Diesel 4 20 Diesel 4 9.4Diesel 5 20 Diesel 5 9.4Diesel 6 30 Diesel 6 30Diesel 7 22.5 Diesel 7 18Diesel 8 22.5 Diesel 8 18Diesel 9 22.5 Diesel 9 18

Diesel 10 22.5 Gas 1 32.42Diesel 11 22.5 Gas 2 40.93

Gas 1 27.24 Gas Mobile 1 15Gas 2 40.93 - -

Peak power demand in 2013 was 251 MW, and it is expected to reach 384 MW by 2020.At present, the power system is operated at 66 kV in the transmission network and at 20 kV at

the distribution level. The transmission network has eight buses. The rated voltage of the submarinecable is 66 kV, with a transmission capacity of 60 MVA.

New substations and transmission lines of 132-kV rated voltage have been planned for 2020. Inaddition, the undersea link will be reinforced by a second submarine cable of 132-kV rated voltageand 130 MVA of capacity. The power system will then have 26 buses at two voltage levels, 66 kV and132 kV.

Currently, the two-island system has five wind farms. Total wind power capacity is 21.86 MW.The most widely-used type of wind turbine is the SCIG with a gear box.

12671

Energies 2015, 8, 12669–12684

The installed wind power is expected to have increased by 2020. Wind power could rise to162 MW [24]. Most of the new wind farms will be full converter wind turbines.

4. System Modeling

This section describes the model of the Lanzarote-Fuerteventura 2020 power system that wasimplemented in PSSrE v32.

Conventional units in power plants, wind power energy converters, power substations and linesof the transmission network have been modeled. The distribution system and power demands arerepresented through loads at corresponding buses.

A single-line diagram of the Lanzarote-Fuerteventura 2020 power system is shown in Figure 1.Transmission lines have been modeled using a π model with their conductances neglected. The

power factor of loads is 0.9. A two-winding transformer model is used for transformers of generationunits and for substation transformers. Resistance and reactance are considered.

Energies 2015, 8, page–page

4

4. System Modeling

This section describes the model of the Lanzarote‐Fuerteventura 2020 power system that was

implemented in PSS®E v32.

Conventional units in power plants, wind power energy converters, power substations and lines

of the transmission network have been modeled. The distribution system and power demands are

represented through loads at corresponding buses.

A single‐line diagram of the Lanzarote‐Fuerteventura 2020 power system is shown in Figure 1.

Transmission lines have been modeled using a π model with their conductances neglected. The

power factor of loads is 0.9. A two‐winding transformer model is used for transformers of generation

units and for substation transformers. Resistance and reactance are considered.

Figure 1. Lanzarote‐Fuerteventura power system expected by the year 2020.

4.1. Conventional Generation

Diesel generators are represented by GENSAL models, and gas turbine generators are modeled

with GENROE. Both models belong to the PSS®E library.

The turbine governor model used for diesel units is DEGOV1 (Woodward diesel governor). The

GAST2A turbine governor model is used for gas turbine units.

PSS®E models used for excitation control in diesel units and gas turbine units are simplified

excitation system (SEXS) and ESDCA1 (IEEE Type DC1A excitation system), respectively.

Figure 1. Lanzarote-Fuerteventura power system expected by the year 2020.

4.1. Conventional Generation

Diesel generators are represented by GENSAL models, and gas turbine generators are modeledwith GENROE. Both models belong to the PSSrE library.

The turbine governor model used for diesel units is DEGOV1 (Woodward diesel governor). TheGAST2A turbine governor model is used for gas turbine units.

PSSrE models used for excitation control in diesel units and gas turbine units are simplifiedexcitation system (SEXS) and ESDCA1 (IEEE Type DC1A excitation system), respectively.

12672

Energies 2015, 8, 12669–12684

4.2. Wind Conversion Energy System

4.2.1. Squirrel Cage Induction Generator (SCIG) Wind Turbine

A SCIG wind turbine, or so-called Type 1, is connected to the grid through a transformer. Thegenerator operates at a nearly fixed speed, related to the grid frequency. The wind turbine generatesactive power when the shaft speed is higher than the grid frequency. The generator consumes reactivepower to create its magnetic field. Thereby, a capacitor bank is used for power factor correction.Normally, in this wind turbine type, there is a soft-starter to limit the higher starting currents [25]. Adiagram of this wind turbine is depicted in Figure 2.


5



A SCIG wind turbine, or so‐called Type 1, is connected to the grid through a transformer. The

generator operates at a nearly fixed speed, related to the grid frequency. The wind turbine generates

active power when the shaft speed is higher than the grid frequency. The generator consumes reactive

power to create its magnetic field. Thereby, a capacitor bank is used for power factor correction.

Normally, in this wind turbine type, there is a soft‐starter to limit the higher starting currents [25]. A

diagram of this wind turbine is depicted in Figure 2.

Figure 2. Diagram of the squirrel cage induction generator (SCIG) wind turbine.

The model used for SCIG wind turbines is the well‐known PSS®E WT1 model [26]. This model

includes the WT1G generator model, the WT12T wind turbine model and the WT12A

pseudo‐governor model.

The WT1G model is a modification of the standard induction machine model and considers rotor

flux dynamics. The WT12T model is based on a two‐mass representation of the wind turbine drive

shaft. It includes rotor blades, the shaft and the machine with the gear box. This model determines

speed deviation. The WT12A model calculates the mechanical torque of the blades’ shaft by

processing rotor speed deviation and active power at generator terminals. A connectivity diagram of

these models can be seen in [26].

4.2.2. Full Converter Wind Turbine

In a full converter wind turbine, a generator is connected to the grid through a power converter.

This allows variable speed in the turbine shaft. The converter rectifies the variable‐frequency AC

power from the generator into DC and then inverts the DC power to AC at the grid frequency. Figure

3 shows this type of wind turbine. The generator can be a wound rotor synchronous generator

(WRSG), a permanent magnet synchronous generator (PMSG) or an SCIG [25]. In this paper, the

WRSG is considered.

Figure 3. Diagram of a full converter wind turbine.


The model used for SCIG wind turbines is the well-known PSSrE WT1 model [26]. Thismodel includes the WT1G generator model, the WT12T wind turbine model and the WT12Apseudo-governor model.

The WT1G model is a modification of the standard induction machine model and considersrotor flux dynamics. The WT12T model is based on a two-mass representation of the wind turbinedrive shaft. It includes rotor blades, the shaft and the machine with the gear box. This modeldetermines speed deviation. The WT12A model calculates the mechanical torque of the blades’ shaftby processing rotor speed deviation and active power at generator terminals. A connectivity diagramof these models can be seen in [26].


In a full converter wind turbine, a generator is connected to the grid through a power converter.This allows variable speed in the turbine shaft. The converter rectifies the variable-frequency ACpower from the generator into DC and then inverts the DC power to AC at the grid frequency.Figure 3 shows this type of wind turbine. The generator can be a wound rotor synchronous generator(WRSG), a permanent magnet synchronous generator (PMSG) or an SCIG [25]. In this paper, theWRSG is considered.


5



A SCIG wind turbine, or so‐called Type 1, is connected to the grid through a transformer. The

generator operates at a nearly fixed speed, related to the grid frequency. The wind turbine generates

active power when the shaft speed is higher than the grid frequency. The generator consumes reactive

power to create its magnetic field. Thereby, a capacitor bank is used for power factor correction.

Normally, in this wind turbine type, there is a soft‐starter to limit the higher starting currents [25]. A

diagram of this wind turbine is depicted in Figure 2.


The model used for SCIG wind turbines is the well‐known PSS®E WT1 model [26]. This model

includes the WT1G generator model, the WT12T wind turbine model and the WT12A

pseudo‐governor model.

The WT1G model is a modification of the standard induction machine model and considers rotor

flux dynamics. The WT12T model is based on a two‐mass representation of the wind turbine drive

shaft. It includes rotor blades, the shaft and the machine with the gear box. This model determines

speed deviation. The WT12A model calculates the mechanical torque of the blades’ shaft by

processing rotor speed deviation and active power at generator terminals. A connectivity diagram of

these models can be seen in [26].


In a full converter wind turbine, a generator is connected to the grid through a power converter.

This allows variable speed in the turbine shaft. The converter rectifies the variable‐frequency AC

power from the generator into DC and then inverts the DC power to AC at the grid frequency. Figure

3 shows this type of wind turbine. The generator can be a wound rotor synchronous generator

(WRSG), a permanent magnet synchronous generator (PMSG) or an SCIG [25]. In this paper, the

WRSG is considered.

Figure 3. Diagram of a full converter wind turbine. Figure 3. Diagram of a full converter wind turbine.

12673

Energies 2015, 8, 12669–12684

Model characteristics of the considered wind turbine are described in [27,28].The dynamic behavior of the wind turbine is related to the converter control system for the

simulation times normally used. Therefore, modeling the converter control system is enough for windturbine representation in almost all cases. Figure 4 shows a diagram for the model of the wind turbine.


6

Model characteristics of the considered wind turbine are described in [27,28].

The dynamic behavior of the wind turbine is related to the converter control system for the

simulation times normally used. Therefore, modeling the converter control system is enough for

wind turbine representation in almost all cases. Figure 4 shows a diagram for the model of the

wind turbine.

Figure 4. Simplified diagram of the full converter wind turbine model.

The power converter allows supporting grid voltage by generating or consuming reactive

power. Reactive power control is available for all of the active power operation range. This can be

seen in Figure 5.

Figure 5. P‐Q characteristic of the full converter wind turbine.

A full converter wind turbine with fault ride‐through capability can remain in operation when

a voltage dip or an overvoltage occurs. The wind turbine is also able to inject or consume additional

reactive power to participate in network voltage restoration.

The reference of total current related to reactive power , is obtained by Equation (1):

, ∆ (1)

P(pu)

Q(pu)

1.0

Q generated Q consumed

0-0.75 0.75


The power converter allows supporting grid voltage by generating or consuming reactive power.Reactive power control is available for all of the active power operation range. This can be seen inFigure 5.


6

Model characteristics of the considered wind turbine are described in [27,28].

The dynamic behavior of the wind turbine is related to the converter control system for the

simulation times normally used. Therefore, modeling the converter control system is enough for

wind turbine representation in almost all cases. Figure 4 shows a diagram for the model of the

wind turbine.


The power converter allows supporting grid voltage by generating or consuming reactive

power. Reactive power control is available for all of the active power operation range. This can be

seen in Figure 5.

Figure 5. P‐Q characteristic of the full converter wind turbine.

A full converter wind turbine with fault ride‐through capability can remain in operation when

a voltage dip or an overvoltage occurs. The wind turbine is also able to inject or consume additional

reactive power to participate in network voltage restoration.

The reference of total current related to reactive power , is obtained by Equation (1):

, ∆ (1)

P(pu)

Q(pu)

1.0

Q generated Q consumed

0-0.75 0.75

Figure 5. P-Q characteristic of the full converter wind turbine.

A full converter wind turbine with fault ride-through capability can remain in operation whena voltage dip or an overvoltage occurs. The wind turbine is also able to inject or consume additionalreactive power to participate in network voltage restoration.

12674

Energies 2015, 8, 12669–12684

The reference of total current related to reactive power IQ,ref is obtained by Equation (1):

IQ,ref “QrefU

` ∆IQ (1)

where Qref is the generated or consumed reactive power of the wind turbine before fault, U is thevoltage at the wind turbine terminals and ∆IQ is the additional reactive current determined as shownin Figure 6.


7

where is the generated or consumed reactive power of the wind turbine before fault, is the

voltage at the wind turbine terminals and ∆ is the additional reactive current determined as shown

in Figure 6.

Figure 6. Additional current related to the reactive power characteristic.

4.2.3. Inertia Emulation

A kind of active power control that some wind turbines can provide is inertia emulation, also

called synthetic inertia or simulated inertia. A wind turbine with inertia emulation capability is able

to increase the generated active power from kinetic energy stored in the rotating mass, by means of

the power converter control. If there is a low system frequency value, the wind turbine injects extra

active power into the network, emulating conventional synchronous generators’ inertia [12,29].

Because kinetic energy is used to increase the active power, the speed of the wind turbine

decreases. In order not to reduce wind turbine speed excessively, emulation inertia can only be

provided during a few seconds. According to some manufacturers, wind turbines can provide this

extra active power within the first 10 s [30–32]. After the power increases, a period of time is needed

to restore the speed to an acceptable value. This recovery period is about twice the power increase

time [16,33].

Depending on the wind turbine, extra active power can vary from 4% to 10% of the wind turbine

rated power [18,34]. Extra active power can be obtained by Equation (2) used in [30].

12

ω12

ω (2)

where is the extra active power duration, is the total moment of inertia of the wind turbine,

ω is the initial rotor angular speed and ω is the rotor angular speed at time .

In this paper, the inertia emulation control model for the full converter wind turbine considered

is based on [16]. This can be seen in Figure 7. The output value depends on the actual value of

the system frequency, as shown in Figure 8.

Figure 6. Additional current related to the reactive power characteristic.

4.2.3. Inertia Emulation

A kind of active power control that some wind turbines can provide is inertia emulation, alsocalled synthetic inertia or simulated inertia. A wind turbine with inertia emulation capability is ableto increase the generated active power from kinetic energy stored in the rotating mass, by means ofthe power converter control. If there is a low system frequency value, the wind turbine injects extraactive power into the network, emulating conventional synchronous generators’ inertia [12,29].

Because kinetic energy is used to increase the active power, the speed of the wind turbinedecreases. In order not to reduce wind turbine speed excessively, emulation inertia can only beprovided during a few seconds. According to some manufacturers, wind turbines can provide thisextra active power within the first 10 s [30–32]. After the power increases, a period of time is neededto restore the speed to an acceptable value. This recovery period is about twice the power increasetime [16,33].

Depending on the wind turbine, extra active power can vary from 4% to 10% of the wind turbinerated power [18,34]. Extra active power Pext can be obtained by Equation (2) used in [30].

Pextt “12

JWTω2r0 ´

12

JWTω2rt (2)

where t is the extra active power duration, JWT is the total moment of inertia of the wind turbine,ωr0

is the initial rotor angular speed andωrt is the rotor angular speed at time t.In this paper, the inertia emulation control model for the full converter wind turbine considered

is based on [16]. This can be seen in Figure 7. The output value Pext depends on the actual value ofthe system frequency, as shown in Figure 8.

12675

Energies 2015, 8, 12669–12684Energies 2015, 8, page–page

8

Figure 7. Basic inertia emulation model block diagram.

Figure 8. Graph behavior of the inertia emulation considered.

is 10% of rated power; fDeadband is 49.85 Hz; and fmin is 49.25 Hz. The recovery period

implemented in the model is twice the power increase time. Inertia emulation is available from 4% of

rated power, and its time response is within 800 ms.

4.3. Protections Relays

Protection relays have been modeled for both the conventional units and the wind turbine

generators. The protection relays considered are: under voltage and overvoltage, under frequency

and over frequency, over current and over speed. Settings for these protection relays are given in

Table 2.

Table 2. Protection relay settings for conventional units and wind turbines.

Generator type

Parameter Under voltage

Overvoltage Under

frequency Over

frequency Over

current 1 Over

current 2 Over speed

Conventional generator

Value 0.75 pu 1.12 pu 47 Hz 52.2 Hz 1.25 pu 3 pu 1.08 pu Delay (s) 0.8 1 1.25 2 6.5 0.5 Instant

Wind turbine Value 0.8 pu 1.12 pu 47 Hz 51 Hz 1.25 pu 3 pu

1.08 pu (Only SCIG WT)

Delay (s) 5 0.3 1.3 0.1 6.5 0.5 Instant (Only SCIG WT)

4.4. Load Shedding Scheme

A load shedding scheme has been implemented in the power system model. Load shedding

relays are associated with load buses. Three steps have been included in the load shedding scheme.

The settings’ frequency values for first, second and third steps are 49.0, 48.9 and 48.8 Hz.



8



is 10% of rated power; fDeadband is 49.85 Hz; and fmin is 49.25 Hz. The recovery period

implemented in the model is twice the power increase time. Inertia emulation is available from 4% of

rated power, and its time response is within 800 ms.


Protection relays have been modeled for both the conventional units and the wind turbine

generators. The protection relays considered are: under voltage and overvoltage, under frequency

and over frequency, over current and over speed. Settings for these protection relays are given in

Table 2.


Generator type

Parameter Under voltage

Overvoltage Under

frequency Over

frequency Over

current 1 Over

current 2 Over speed

Conventional generator

Value 0.75 pu 1.12 pu 47 Hz 52.2 Hz 1.25 pu 3 pu 1.08 pu Delay (s) 0.8 1 1.25 2 6.5 0.5 Instant

Wind turbine Value 0.8 pu 1.12 pu 47 Hz 51 Hz 1.25 pu 3 pu

1.08 pu (Only SCIG WT)

Delay (s) 5 0.3 1.3 0.1 6.5 0.5 Instant (Only SCIG WT)


A load shedding scheme has been implemented in the power system model. Load shedding

relays are associated with load buses. Three steps have been included in the load shedding scheme.

The settings’ frequency values for first, second and third steps are 49.0, 48.9 and 48.8 Hz.


Pext max is 10% of rated power; f Deadband is 49.85 Hz; and f min is 49.25 Hz. The recovery periodimplemented in the model is twice the power increase time. Inertia emulation is available from 4% ofrated power, and its time response is within 800 ms.


Protection relays have been modeled for both the conventional units and the wind turbinegenerators. The protection relays considered are: under voltage and overvoltage, under frequencyand over frequency, over current and over speed. Settings for these protection relays are given inTable 2.


Generatortype Parameter Under

voltage Overvoltage Underfrequency

Overfrequency

Overcurrent 1

Overcurrent 2 Over speed

Conventionalgenerator

Value 0.75 pu 1.12 pu 47 Hz 52.2 Hz 1.25 pu 3 pu 1.08 pu

Delay (s) 0.8 1 1.25 2 6.5 0.5 Instant

Wind turbine

Value 0.8 pu 1.12 pu 47 Hz 51 Hz 1.25 pu 3 pu 1.08 pu (OnlySCIG WT)

Delay (s) 5 0.3 1.3 0.1 6.5 0.5 Instant (OnlySCIG WT)


A load shedding scheme has been implemented in the power system model. Load sheddingrelays are associated with load buses. Three steps have been included in the load shedding scheme.The settings’ frequency values for first, second and third steps are 49.0, 48.9 and 48.8 Hz.

12676

Energies 2015, 8, 12669–12684

Tripping of the first and second steps causes a load shedding higher than 10%. The setting timedelay for this tripping is 0.45 s.

5. Results and Discussion

Simulation results have been analyzed to investigate which of the three criteria mentioned inthe Introduction section determines the CCT in the buses of the Lanzarote-Fuerteventura 2020 powersystem. The answer was found to be load shedding in all cases. Figure 9 presents the CCT valuesobtained for all of the wind turbine types: SCIG, full converter and full converter with inertiaemulation capability. It can be seen that there is a progressive decrease in all of them when windpower is injected for the five buses under short circuit.


9

Tripping of the first and second steps causes a load shedding higher than 10%. The setting time

delay for this tripping is 0.45 s.

5. Results and Discussion

Simulation results have been analyzed to investigate which of the three criteria mentioned in the

Introduction section determines the CCT in the buses of the Lanzarote‐Fuerteventura 2020 power

system. The answer was found to be load shedding in all cases. Figure 9 presents the CCT values

obtained for all of the wind turbine types: SCIG, full converter and full converter with inertia

emulation capability. It can be seen that there is a progressive decrease in all of them when wind

power is injected for the five buses under short circuit.

(a) (b)

(c) (d)

(e)

Figure 9. Critical clearing time (CCT) values obtained: (a) Punta Grande bus; (b) Haría‐Teguise bus;

(c) Las Salinas bus; (d) Jandía bus; (e) Corralejo bus.

100

120

140

160

180

200

220

240

260

0 20 40 60 80 100 120 140

Critical Clearing Time (m

s)

Wind Power Generated (MW)

Punta Grande Bus

150

200

250

300

350

400

450

500

550

0 20 40 60 80 100 120 140 160


s)


Haría‐Teguise Bus

150

170

190

210

230

250

270

290

310

330

0 20 40 60 80 100 120 140


s)


Las Salinas Bus

150

250

350

450

550

650

750

850

0 20 40 60 80 100 120 140


s)


Jandía Bus

150

200

250

300

350

400

450

500

550

600

0 20 40 60 80 100 120 140


s)


Corralejo Bus

Full Converter Wind Turbine

Full Converter Wind Turbinewith Iner a Emula on

SCIG Wind Turbine

Figure 9. Critical clearing time (CCT) values obtained: (a) Punta Grande bus; (b) Haría-Teguise bus;(c) Las Salinas bus; (d) Jandía bus; (e) Corralejo bus.

12677

Energies 2015, 8, 12669–12684

The decrease in the CCT values is due to the outage of wind farms. A three-phase short circuitcauses a voltage dip. An overvoltage can also take place after the fault is cleared, which can be seen inothers papers dealing with small isolated power systems [35]. Because of the small size of the studiedpower system, voltage dip and subsequent overvoltage can have an impact on the overall powersystem. This subsequent overvoltage causes the outage of the wind farms after the short circuit iscleared. The outages can be seen in Figure 10, which includes an example of a short circuit for everystudied case.


10

The decrease in the CCT values is due to the outage of wind farms. A three‐phase short circuit

causes a voltage dip. An overvoltage can also take place after the fault is cleared, which can be seen

in others papers dealing with small isolated power systems [35]. Because of the small size of the

studied power system, voltage dip and subsequent overvoltage can have an impact on the overall

power system. This subsequent overvoltage causes the outage of the wind farms after the short circuit

is cleared. The outages can be seen in Figure 10, which includes an example of a short circuit for every

studied case.

(a) (b)

(c)

Frequency - 20MW Wind Power





Figure 10. Time evolution of voltage and frequency for a three-phase short circuit at: (a) Las Salinassubstation, with 260 ms of clearing time and an SCIG wind turbine; (b) Punta Grande substation,with 230 ms of clearing time and a full converter wind turbine; and (c) Haría-Teguise, with 420 ms ofclearing time with a full converter wind turbine and with inertia emulation capability.

12678

Energies 2015, 8, 12669–12684

The swing equation of the synchronous machine (Equation (3)) is related to power systemstability, and it can be used to study how wind farms have an impact on the power system.

dωdt

“1

2HpPm ´ Peq (3)

where dωdt is the rotor acceleration, H is the inertia constant, Pm is the mechanical power and Pe is the

electrical power [36].Usually, when a short circuit occurs, electrical power at generator terminals Pe diminishes. This

causes a great acceleration in conventional generators dωdt and system frequency rises.

When the short circuit is cleared, electrical power starts to retrieve its previous value. Moreover,turbine governors of the generators have been actuated to reduce mechanical power Pm. Both effectscause the decrease of dω

dt , and the frequency value drops below the rated value.Wind farm outages take place when the frequency is dropping. This leads to a greater imbalance

between Pe and Pm, followed by higher frequency deviations, as can be seen in Figure 10.These frequency deviations can reach 48.9 Hz, causing the tripping of the first and second step

of load shedding relays. The load shedding is greater than 10% of the total load and defines the CCTvalue. Thereby, the more the wind power is disconnected, the more CCT decreases.

Some wind farm outages caused by over frequency values can also be seen. These wind farmoutages due to over frequency are about 13% of total studied short circuit cases.

The decrease in the number of running conventional generators also causes the decrease in theCCT values. In order to fulfill the Boucherot theorem, progressive introduction of wind power intothe power system can require the disconnection of some conventional generators. This reduction inthe number of conventional units causes a decrease in the total power system inertia. Figure 11 showsthe inertia constant of the power system for the wind power amounts considered.


11

Figure 10. Time evolution of voltage and frequency for a three‐phase short circuit at: (a) Las Salinas

substation, with 260 ms of clearing time and an SCIG wind turbine; (b) Punta Grande substation, with

230 ms of clearing time and a full converter wind turbine; and (c) Haría‐Teguise, with 420 ms of

clearing time with a full converter wind turbine and with inertia emulation capability.

The swing equation of the synchronous machine (Equation (3)) is related to power system

stability, and it can be used to study how wind farms have an impact on the power system.

dωd

12

(3)

where is the rotor acceleration, H is the inertia constant, is the mechanical power and is

the electrical power [36].

Usually, when a short circuit occurs, electrical power at generator terminals diminishes. This

causes a great acceleration in conventional generators and system frequency rises.

When the short circuit is cleared, electrical power starts to retrieve its previous value. Moreover,

turbine governors of the generators have been actuated to reduce mechanical power . Both effects

cause the decrease of , and the frequency value drops below the rated value.

Wind farm outages take place when the frequency is dropping. This leads to a greater imbalance

between and , followed by higher frequency deviations, as can be seen in Figure 10.

These frequency deviations can reach 48.9 Hz, causing the tripping of the first and second step

of load shedding relays. The load shedding is greater than 10% of the total load and defines the CCT

value. Thereby, the more the wind power is disconnected, the more CCT decreases.

Some wind farm outages caused by over frequency values can also be seen. These wind farm

outages due to over frequency are about 13% of total studied short circuit cases.

The decrease in the number of running conventional generators also causes the decrease in the

CCT values. In order to fulfill the Boucherot theorem, progressive introduction of wind power into

the power system can require the disconnection of some conventional generators. This reduction in

the number of conventional units causes a decrease in the total power system inertia. Figure 11 shows

the inertia constant of the power system for the wind power amounts considered.

Figure 11. Power system inertia for different amounts of wind power.

Considering again the swing equation, a lower inertia constant H causes a higher frequency

variation when an imbalance between and occurs.

Moreover, the reduction of active power‐frequency control and of voltage control capabilities is

a consequence of the decrease in the number of conventional generator units. The decrease of these

capabilities also causes a CCT decrease.

Figure 9 also shows different evolutions of CCT for different wind turbines. The CCT values

obtained for full converter wind turbines are higher than CCT values for SCIG wind turbines, except

in the Jandía bus, which is located at the end of a long transmission line.

The Jandía bus is located far from the nearest power plant. Besides, there are wind farms

connected at several buses at different points of this transmission line. This can be seen in Figure 1.

Figure 11. Power system inertia for different amounts of wind power.

Considering again the swing equation, a lower inertia constant H causes a higher frequencyvariation when an imbalance between Pe and Pm occurs.

Moreover, the reduction of active power-frequency control and of voltage control capabilities isa consequence of the decrease in the number of conventional generator units. The decrease of thesecapabilities also causes a CCT decrease.

Figure 9 also shows different evolutions of CCT for different wind turbines. The CCT valuesobtained for full converter wind turbines are higher than CCT values for SCIG wind turbines, exceptin the Jandía bus, which is located at the end of a long transmission line.

The Jandía bus is located far from the nearest power plant. Besides, there are wind farmsconnected at several buses at different points of this transmission line. This can be seen in Figure 1.When a short circuit takes place at the Jandía bus, full converter wind turbines contribute to obtaininghigher voltages than in the case of SCIG. Therefore, loads at these buses are higher and cause a deeperfrequency fall.

12679

Energies 2015, 8, 12669–12684

The higher CCT values with full converter wind turbines are due to the fault ride throughcapability of the full converter wind turbines. An example of the performance of this capability isshown in Figure 12.


12

When a short circuit takes place at the Jandía bus, full converter wind turbines contribute to obtaining

higher voltages than in the case of SCIG. Therefore, loads at these buses are higher and cause a deeper

frequency fall.

The higher CCT values with full converter wind turbines are due to the fault ride through

capability of the full converter wind turbines. An example of the performance of this capability is

shown in Figure 12.

Figure 12. Reactive power generated or consumed by wind farm Number 2 in both the SCIG wind

turbine case and the full converter wind turbine case, for a 300‐ms three‐phase short circuit.

Figure 12 shows that SCIG tripped due to an overvoltage condition at 2.4 s, causing the reactive

power changes to zero at 2.4 s. In the case of the full power converter, what the figure shows is the

network voltage support obtained by means of generating or consuming reactive power (Figures 5

and 6). After the short circuit is cleared at 1.3 s, voltage recovery is slow, and the converter continues

generating reactive power to support the bus voltage. Later, between 2.1 s and 2.7 s, wind generators

with a voltage support capability absorb the reactive power to avoid a transient overvoltage

condition that could take place as a part of the voltage recovery evolution. In Figure 10a, a similar

transient overvoltage is shown for the SCIG case

The advantages of the voltage support capability can also be seen in Figure 13. Voltages have a

higher minimum value and a lower maximum value in the case of full converter wind turbines,

avoiding tripping by means of protective relay operation.

Figure 12. Reactive power generated or consumed by wind farm Number 2 in both the SCIG windturbine case and the full converter wind turbine case, for a 300-ms three-phase short circuit.

Figure 12 shows that SCIG tripped due to an overvoltage condition at 2.4 s, causing the reactivepower changes to zero at 2.4 s. In the case of the full power converter, what the figure showsis the network voltage support obtained by means of generating or consuming reactive power(Figures 5 and 6). After the short circuit is cleared at 1.3 s, voltage recovery is slow, and the convertercontinues generating reactive power to support the bus voltage. Later, between 2.1 s and 2.7 s,wind generators with a voltage support capability absorb the reactive power to avoid a transientovervoltage condition that could take place as a part of the voltage recovery evolution. In Figure 10a,a similar transient overvoltage is shown for the SCIG case

The advantages of the voltage support capability can also be seen in Figure 13. Voltages havea higher minimum value and a lower maximum value in the case of full converter wind turbines,avoiding tripping by means of protective relay operation.


13

Figure 13. Time evolution of voltage and frequency for a 300‐ms three‐phase short circuit at the Las

Salinas bus with 80 MW of total wind power.

On the contrary, SCIG tripping due to that transient overvoltage condition leads to an imbalance

between generated power and load for the whole power system, causing a larger system frequency

deviation and the subsequent load shedding.

Inertia emulation capability does not influence reactive power generated or consumed by full

converter wind turbines when a fault takes place. Therefore, a curve for full converter wind turbines

with inertia emulation capability was not included in Figure 12.

Since full converter wind turbines generate additional reactive power during a short circuit and

this improves voltage values, the speed‐up of the conventional generators is lower with full converter

wind turbines and leads to lower frequency deviations. This can be seen in Figure 13, which shows

voltage, as well as the system frequency response. This figure shows that the maximum frequency

value reached is lower for full converter wind turbines as a consequence of the higher voltage values

during the short circuit.

Moreover, since a full converter wind turbine consumes additional reactive power when the

subsequent overvoltage takes place, this kind of wind turbine could avoid its own overvoltage outage

in some short circuit cases. In these cases, such as in Figure 13, the frequency values obtained do not

cause tripping of the first and second step of load shedding relays.

Figure 9 shows differences between CCT values for full converter wind turbines with and

without inertia emulation capability. CCT values for full converter wind turbines with inertia

emulation capability are about 3.6% higher than those obtained without inertia emulation capability.

This is due to the power increase produced by inertia emulation. When the frequency starts to

fall, a wind turbine with inertia emulation capability injects extra active power, as can be seen in the

example presented in Figure 14.

Figure 13. Time evolution of voltage and frequency for a 300-ms three-phase short circuit at the LasSalinas bus with 80 MW of total wind power.

12680

Energies 2015, 8, 12669–12684

On the contrary, SCIG tripping due to that transient overvoltage condition leads to an imbalancebetween generated power and load for the whole power system, causing a larger system frequencydeviation and the subsequent load shedding.

Inertia emulation capability does not influence reactive power generated or consumed by fullconverter wind turbines when a fault takes place. Therefore, a curve for full converter wind turbineswith inertia emulation capability was not included in Figure 12.

Since full converter wind turbines generate additional reactive power during a short circuit andthis improves voltage values, the speed-up of the conventional generators is lower with full converterwind turbines and leads to lower frequency deviations. This can be seen in Figure 13, which showsvoltage, as well as the system frequency response. This figure shows that the maximum frequencyvalue reached is lower for full converter wind turbines as a consequence of the higher voltage valuesduring the short circuit.

Moreover, since a full converter wind turbine consumes additional reactive power when thesubsequent overvoltage takes place, this kind of wind turbine could avoid its own overvoltage outagein some short circuit cases. In these cases, such as in Figure 13, the frequency values obtained do notcause tripping of the first and second step of load shedding relays.

Figure 9 shows differences between CCT values for full converter wind turbines with andwithout inertia emulation capability. CCT values for full converter wind turbines with inertiaemulation capability are about 3.6% higher than those obtained without inertia emulation capability.

This is due to the power increase produced by inertia emulation. When the frequency starts tofall, a wind turbine with inertia emulation capability injects extra active power, as can be seen in theexample presented in Figure 14.Energies 2015, 8, page–page

14

Figure 14. Active power generated by wind farm Number 2 for both the full converter wind turbine

and the full converter with inertia emulation capability, for a 300‐ms three‐phase short circuit at the

Las Salinas bus with 80 MW of total wind power

Figure 13 shows how this extra active power reduces the frequency drop. In this way, lower

frequency deviation and higher frequency nadir values are produced. For instance, the frequency

nadir is about 10.4% higher when a 150‐ms three‐phase short circuit occurs at the five analyzed buses.

If CCT values for the full converter wind turbine with inertia emulation capability are compared

to corresponding values for an SCIG wind turbine, the average value is 24.3% higher, and the

frequency nadir is 34.8% higher.

Thereby, inertia emulation is able to improve the CCT values in the simulations performed.

However, inertia emulation could not avoid load shedding when a long short circuit caused severe

frequency deviations.

6. Conclusions

This paper analyzes the effects of wind power on CCT in small and isolated power systems.

CCT is a parameter used for protection scheme design and is therefore related to power system

security.

Three types of wind turbines are studied: the SCIG, the full converter and the full converter with

inertia emulation capability.

The analysis shows that load shedding can be caused by large frequency deviations. This criteria

determines the CCT values in the Lanzarote‐Fuerteventura power system. A decrease in CCT values

can be seen when wind power is increased for all types of wind turbines studied. This is due to the

wind farm outages caused by overvoltage in small and isolated power systems. Thereby, voltage time

evolution can have an impact on CCT values in such systems. Reduction of the inertia and power

frequency control and voltage control capabilities in these power systems also causes a decrease in

CCT values.

Results also show that the behavior of each particular type of wind turbine establishes

differences in CCT values. CCT values for a full converter are higher than those obtained for an SCIG.

When a short circuit takes place, full converter wind turbines remain in operation and generate or

consume reactive power. Therefore, a full converter wind turbine could avoid its own overvoltage

outage, and this leads to higher CCT values than in the case of an SCIG wind turbine.

The highest CCT values are obtained for the full converter wind turbine with inertia emulation

capability. These CCT values are 3.6% higher than those for full converter wind turbines without this

capability. The percentage rises to 34.8% if these results are compared with the results for SCIG wind

turbines. Extra active power generated by full converter wind turbines with inertia emulation

achieves lower system frequency deviation and implies the higher CCT values mentioned above.

Thereby, inertia emulation can improve the CCT values in a small and isolated power system.

Therefore, these issues suggest the existence of a relationship between wind power and CCT.

Hence, it is necessary to consider these aspects in studies for CCT determination or studies of wind

power integration in small and isolated power systems.

Figure 14. Active power generated by wind farm Number 2 for both the full converter wind turbineand the full converter with inertia emulation capability, for a 300-ms three-phase short circuit at theLas Salinas bus with 80 MW of total wind power

Figure 13 shows how this extra active power reduces the frequency drop. In this way, lowerfrequency deviation and higher frequency nadir values are produced. For instance, the frequencynadir is about 10.4% higher when a 150-ms three-phase short circuit occurs at the five analyzed buses.

If CCT values for the full converter wind turbine with inertia emulation capability are comparedto corresponding values for an SCIG wind turbine, the average value is 24.3% higher, and thefrequency nadir is 34.8% higher.

Thereby, inertia emulation is able to improve the CCT values in the simulations performed.However, inertia emulation could not avoid load shedding when a long short circuit caused severefrequency deviations.

6. Conclusions

This paper analyzes the effects of wind power on CCT in small and isolated power systems.

12681

Energies 2015, 8, 12669–12684

CCT is a parameter used for protection scheme design and is therefore related to powersystem security.

Three types of wind turbines are studied: the SCIG, the full converter and the full converter withinertia emulation capability.

The analysis shows that load shedding can be caused by large frequency deviations. This criteriadetermines the CCT values in the Lanzarote-Fuerteventura power system. A decrease in CCT valuescan be seen when wind power is increased for all types of wind turbines studied. This is due tothe wind farm outages caused by overvoltage in small and isolated power systems. Thereby, voltagetime evolution can have an impact on CCT values in such systems. Reduction of the inertia and powerfrequency control and voltage control capabilities in these power systems also causes a decrease inCCT values.

Results also show that the behavior of each particular type of wind turbine establishesdifferences in CCT values. CCT values for a full converter are higher than those obtained for anSCIG. When a short circuit takes place, full converter wind turbines remain in operation and generateor consume reactive power. Therefore, a full converter wind turbine could avoid its own overvoltageoutage, and this leads to higher CCT values than in the case of an SCIG wind turbine.

The highest CCT values are obtained for the full converter wind turbine with inertia emulationcapability. These CCT values are 3.6% higher than those for full converter wind turbines withoutthis capability. The percentage rises to 34.8% if these results are compared with the results for SCIGwind turbines. Extra active power generated by full converter wind turbines with inertia emulationachieves lower system frequency deviation and implies the higher CCT values mentioned above.Thereby, inertia emulation can improve the CCT values in a small and isolated power system.

Therefore, these issues suggest the existence of a relationship between wind power and CCT.Hence, it is necessary to consider these aspects in studies for CCT determination or studies of windpower integration in small and isolated power systems.

A more precise estimation of CCT values can be performed, and therefore, a more effectivegeneral protection scheme can be achieved to provide security and reliability to the power system.Moreover, in order to increase wind power penetration into small and isolated power systems, someimprovements can be made, such as, for example, adaptation of grid codes or enhancement of reactivepower control capabilities.

Finally, as CCT values are improved by the inertia emulation capability from wind turbines, thiswind power capability could increase the penetration of wind power into small and isolated powersystems. Further analysis with inertia emulation capability is required in order to investigate howmuch wind power penetration can be introduced in each isolated power system.

Acknowledgments: A part of this work was done within the Transition to a Sustainable Energy Model forMadeira, the Azores and the Canary Islands (TRES) project, belonging to the Madeira-Canarias-Azores (MAC2007–2013) Transnational Cooperation Program and funded by the European Union through European RegionalDevelopment Fund (ERDF).

Conflicts of Interest: The authors declare no conflict of interest.

References

1. Global Wind Energy. Council Global Wind Energy Outlook 2014. Available online: http://www.gwec.net/wp-content/uploads/2015/03/GWEC_Global_Wind_2014_Report_LR.pdf (accessed on 6 May 2014).

2. Egido, I.; Fernandez-Bernal, F.; Centeno, P.; Rouco, L. Maximum frequency deviation calculation in smallisolated power systems. IEEE Trans. Power Syst. 2009, 24, 1731–1738. [CrossRef]

3. Lenga, O. Application of critical clearing time as a transient stability criterion for power system expansionplanning. In Proceedings of the 2014 IEEE International Conference on Intelligent Energy and PowerSystems (IEPS), Kyiv, Ukraine, 2–6 June 2014.

4. Shi, L.; Sun, S.; Yao, L.; Ni, Y.; Bazargan, M. Effects of wind generation intermittency and volatility on powersystem transient stability. IET Renew. Power Gener. 2014, 8, 509–521. [CrossRef]

12682

http://dx.doi.org/10.1109/TPWRS.2009.2030399

http://dx.doi.org/10.1049/iet-rpg.2013.0028

Energies 2015, 8, 12669–12684

5. Criterios Generales de Protección del Sistema Eléctrico Peninsular Español. Available online:http://www.ree.es/sites/default/files/downloadable/criterios_proteccion_sistema_1995.pdf (accessedon 30 September 2015). (In Spanish).

6. Criterios Generales de Protección de Los Sistemas Eléctricos Insulares y Extrapeninsulares. Availableonline: http://www.ree.es/sites/default/files/downloadable/criterios_proteccion_sistema_2005_v2.pdf(accessed on 3 October 2014). (In Spanish).

7. Ribot, J.; Vaquero, J.; López, S.; Rodríguez, F.; Martínez, J.L.; Lozano, P.; Fernández, C.; Nicolau, G.;Roca, J.M.; Rodríguez, B.; et al. Evolution in the Spanish Power System: New Protection Coordination Criteriaand Critical Clearing Time Calculation Methodology; International Council on Large Electric Systems (CIGRE):Paris, France, 2006.

8. Metodología de Aplicación de los Criterios Generales de Protección de los Sistemas Eléctricos Insularesy Extrapeninsulares. Available online: http://www.ree.es/sites/default/files/criterios_proteccion_sistema_2007.pdf (accessed on 30 September 2015). (In Spanish).

9. Liu, C.; Chen, Z.; Bak, C.L.; Liu, Z.; Lund, P.; Ronne-Hansen, P. Transient stability assessment of powersystem with large amount of wind power penetration: The Danish case study. In Proceedings of the 2012Conference on Power & Energy IPEC, Ho Chi Minh City, Vietnam, 12–14 December 2012.

10. Chowdhury, M.A.; Hosseinzadeh, N.; Shen, W.X.; Pota, H.R. Comparative study on fault responses ofsynchronous generators and wind turbine generators using transient stability index based on transientenergy function. Int. J. Electr. Power Energy Syst. 2013, 51, 145–152. [CrossRef]

11. Mitra, A.; Chatterjee, D. A sensitivity based approach to assess the impacts of integration of variable speedwind farms on the transient stability of power systems. Renew. Energy 2013, 60, 662–671. [CrossRef]

12. Morren, J.; de Haan, S.W.H.; Kling, W.L.; Ferreira, J.A. Wind turbines emulating inertia and supportingprimary frequency control. IEEE Trans. Power Syst. 2006, 21, 433–434. [CrossRef]

13. Conroy, J.F.; Watson, R. Frequency response capability of full converter wind turbine generators incomparison to conventional generation. IEEE Trans. Power Syst. 2008, 23, 649–656. [CrossRef]

14. Ullah, N.R.; Thiringer, T.; Karlsson, D. Temporary primary frequency control support by variable speedwind turbines—Potential and applications. IEEE Trans. Power Syst. 2008, 23, 601–612. [CrossRef]

15. Keung, P.-K.; Li, P.; Banakar, H.; Ooi, B.T. Kinetic energy of wind-turbine generators for system frequencysupport. IEEE Trans. Power Syst. 2008, 24, 279–287. [CrossRef]

16. Brisebois, J.; Aubut, N. Wind farm inertia emulation to fulfill Hydro-Québec’s specific need. In Proceedingsof the 2011 IEEE Power and Energy Society General Meeting, Detroit, MI, USA, 24–29 July 2011.

17. Björnstedt, J. Integration of Non-synchronous Generation Frequency Dynamics. Ph.D. Thesis, LundUniversity, Lund, Sweden, 2012.

18. Gonzalez-Longatt, F. Impact of synthetic inertia from wind power on the protection/control schemesof future power systems: Simulation study. In Proceedings of the 11th International Conference onDevelopments in Power Systems Protection DPSP 2012, Birmingham, UK, 23–26 April 2012.

19. The utilization of synthetic inertia from wind farms and its impact on existing speed governors andsystem performance. Available online: http://www.elforsk.se/Rapporter/?rid=13_02_ (accessed on 21January 2014).

20. Gevorgian, V.; Zhang, Y.; Ela, E. Investigating the impacts of wind generation participation ininterconnection frequency response. IEEE Trans. Sustain. Energy 2014, 6, 1004–1012. [CrossRef]

21. Papaefthymiou, S.V.; Lakiotis, V.G.; Margaris, I.D.; Papathanassiou, S.A. Dynamic analysis of islandsystems with wind-pumped-storage hybrid power stations. Renew. Energy 2015, 74, 544–554. [CrossRef]

22. Yang, Y.; Wei, H.; Zhao, X.; Zhang, Q. Direct critical clearing time calculation based on interiorpoint nonlinear programming. In Proceedings of the 2012 Asia-Pacific Power and Energy EngineeringConference (APPEEC), Shanghai, China, 27–29 March 2012.

23. DYNA. El Sistema Eléctrico Español en 2013. DYNA Ing. Ind. 2014, 89, 253, (In Spanish).24. Ministerio de Industria, Turismo y Comercio, Gobierno de España. Planificación de los Sectores

de Electricidad y Gas 2012–2020. Desarrollo de las Redes de Transporte. Primer Borrador(July 2011). Available online: http://www.minetur.gob.es/energia/esES/Novedades/Documents/PlanificacionElectricidadGas_2012_2020.pdf (accessed on 2 June 2013). (In Spanish).

12683

http://dx.doi.org/10.1016/j.ijepes.2013.02.025

http://dx.doi.org/10.1016/j.renene.2013.06.002





http://dx.doi.org/10.1109/TSTE.2014.2343836

http://dx.doi.org/10.1016/j.renene.2014.08.062

Energies 2015, 8, 12669–12684

25. Camm, E.H.; Behnke, M.R.; Bolado, O.; Bollen, M.; Bradt, M.; Brooks, C.; Dilling, W.; Edds, M.; Hejdak, W.J.;Houseman, D.; et al. Characteristics of Wind Turbine Generators for Wind Power Plants. In Proceedings ofthe 2009 IEEE Power & Energy Society General Meeting, Calgary, AB, Canada, 26–30 July 2009.

26. Ellis, A.; Muljadi, E.; Sanchez-Gasca, J.; Kazachkov, Y. Generic models for simulation of wind power plantsin bulk system planning studies. In Proceedings of the 2011 IEEE Power and Energy Society GeneralMeeting, Detroit, MI, USA, 24–29 July 2011.

27. Taveira, N.; Mendonça, A.; Quitmann, E.; Moreira, C.L.; Silva, B.; Moreira, J.; Pereira, C. The connectionstudies for 1200 MW wind power integration in Portugal. In Proceedings of the 11th InternationalWorkshop on Large-Scale Integration of Wind Power into Power Systems as well as on TransmissionNetworks for Offshore Wind Power Plants, Lisboa, Portugal, 13–15 November 2012.

28. Fischer, M.; Schellschmidt, M. Fault ride through performance of wind energy converters with FACTScapabilities in response to up-to-date German grid connection requirements. In Proceedings of the 2011IEEE/PES Power Systems Conference and Exposition (PSCE), Phoenix, AZ, USA, 20–23 March 2011.

29. Lalor, G.; Mullane, A.; O’Malley, M. Frequency control and wind turbine technologies. IEEE Trans. PowerSyst. 2005, 20, 1905–1913. [CrossRef]

30. Sun, Y.-Z.; Zhang, Z.-S.; Li, G.-J.; Lin, J. Review on frequency control of power systems with windpower penetration. In Proceedings of the 2010 International Conference on Power System Technology(POWERCON), Hangzhou, Zhejiang, China, 24–28 October 2010.

31. Wachtel, S.; Beekmann, A. Contribution of wind energy converters with inertia emulation to frequencycontrol and frequency stability in power systems. In Proceedings of the 8th International Workshop onLarge Scale Integration of Wind Power into Power Systems as well as on Offshore Wind Farms, Bremen,Germany, 14–15 October 2009; Energynautics: Darmstadt, Germany, 2009.

32. Burra, R.; Ambekar, A.; Narang, H.; Liu, E.; Mehendale, C.; Thirer, L.; Longtin, K.; Shah, M.; Miller, N.GE brilliant wind farms. In Proceedings of the 2014 IEEE Symposium Power Electronics and Machines forWind and Water Applications (PEMWA), Milwaukee, WI, USA, 24–26 July 2014.

33. Taveira, N. Improving the grid compatibility with Enercon wind energy converters. In Proceedings of theConvención Eólica 2010, Madrid, Spain, 15–16 June 2010.

34. Ruttledge, L.; Miller, N.W.; O’Sullivan, J.; Flynn, D. Frequency response of power systems with variablespeed wind turbines. IEEE Trans. Sustain. Energy 2012, 3, 683–691. [CrossRef]

35. Merino, J.; Veganzones, C.; Sánchez, J.A.; Martínez, S.; Platero, C.A. Power system stability of a small sizedisolated network supplied by a combined wind-pumped storage generation system: A case study in theCanary Islands. Energies 2012, 5, 2351–2369. [CrossRef]

36. Anderson, P.; Fouad, A. Power System Control and Stability, 1st ed.; The Iowa State University Press: Ames,IA, USA, 1977.

© 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an openaccess article distributed under the terms and conditions of the Creative Commons byAttribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

12684


http://dx.doi.org/10.1109/TSTE.2012.2202928

http://dx.doi.org/10.3390/en5072351